Plant stages with biotic, indirect defences are more
palatable and suffer less herbivory than their
undefended counterparts
ANA L. LLANDRES
1*, MIGUEL A. RODRÍGUEZ-GIRONÉS
1and RODOLFO DIRZO
21Department of Functional and Evolutionary Ecology, Estación Experimental de Zonas Áridas (CSIC), Carretera de Sacramento s/n, E-04120, La Cañada de San Urbano-Almería, Spain 2Department of Biology, Stanford University, Stanford, CA 94305, USA
Received 29 March 2010; revised 4 June 2010; accepted for publication 4 June 2010bij_1521536..543
Plants have evolved several anti-herbivory strategies, including direct defences, such as mechanical and chemical defences, and indirect or biotic defences, such as the recruitment of defending animals. We examined whether the investment plants make in direct defences differs between those which do and do not invest in biotic defences, by comparing standing herbivory and palatability of congeneric species with and without indirect defences at two ontogenetic stages: before and after the onset of indirect defences. We used Cordia alliodora and Croton suberosus as the species with indirect defences and Cordia elaeagnoides and Croton pseudoniveus as the species without indirect defences. We predicted that herbivores would prefer to eat species and stages with indirect defences to those without them. As predicted, we found that herbivores preferred species and ontogenetic stages with indirect defences in all cases. Overall, however, natural levels of herbivory were lower in species with indirect defences. We conclude that indirect defences offer effective protection against herbivores and posit that their recruitment allows plants to reduce investment in other defence mechanisms. Our results support the notion that plants trade-off between direct and indirect defensive strategies. © 2010 The Linnean Society of London, Biological Journal of the Linnean Society, 2010, 101, 536–543.
ADDITIONAL KEYWORDS: anti-herbivore defences – herbivore preference – Mexico – mutualistic defenders – plant ontogeny – trade-offs – tropical dry forest.
INTRODUCTION
Plants protect themselves against herbivores with a diverse array of mechanisms, including direct and indirect defences. Direct defences are traits that decrease plant quality as food for herbivores, thereby reducing herbivore damage (Gowda et al., 2003; Rohde, Molis & Wahl, 2004). Indirect or biotic defences result from traits that attract animals, which in turn defend the plant from its herbivores (Janzen, 1966, 1969; Schemske, 1980; Domínguez, Dirzo & Bullock, 1989; Di Giusto et al., 2001; Cuautle & Rico-Gray, 2003; Del Val & Dirzo, 2004; Kost & Heil, 2005). Anti-herbivore defences often differ among ontogenetic
stages within plant species and there are two primary hypotheses to explain patterns in this variation. The first explanation is concerned with resource allocation and proposes that young plants are unable to produce high levels of defences because they lack the resources to produce them (Herms & Mattson, 1992). The second explanation, the optimal-defence hypothesis, suggests that younger stages are more vulnerable to herbivore attacks and should therefore be better defended (Bryant et al., 1992). Combining these two explana-tions, Boege & Marquis (2005) predicted a decrease in the levels of defences from seedlings to the juvenile stage and an increase in defences from the juvenile to the mature stage. Some of the patterns predicted by Boege & Marquis (2005) have been supported by a meta-analysis of available studies, but the broader picture is more complex: the relationship between plant age and defence level differs between plant life
histories, herbivore groups and geographical location (Barton & Koricheva, 2010). In particular, for non-boreal woody plants, Barton & Koricheva (2010) detected a general increase in the level of defences from the seedling to the juvenile to the mature stages. From their results, Barton & Koricheva (2010) con-cluded that mechanistic constraints probably deter-mine the defence strategy adopted by each species.
Mechanistic constraints such as resource limitation and physiological trade-offs influence ontogenetic variation in plant defences in many systems. For instance, certain secondary compounds are stored in specialized tissue that is not present in young seed-lings, leading to an ontogenetic increase in the use of these defences when the tissues become available (Goralka & Langenheim, 1996; Goralka, Schumaker & Langenheim, 1996 – cited in Barton & Koricheva, 2010). Likewise, some indirect defences rely on tissues and/or structures that are not present in young seedlings. This led Boege & Marquis (2005) to predict that, unlike direct defences, these indirect defences will increase as plants grow from seedlings to juveniles.
Some plants, known as myrmecophytes (Webber et al., 2007), produce feeding rewards for ants (nectar or nutritious bodies). Frequently, mymecophytes produce specialized structures called domatia, in which ants nest. Domatia may also, in some cases, provide food for the ants (Fiala & Maschwitz, 1992; Gaume, McKey & Terrin, 1998). In turn, ants protect host plants from herbivore attack (Janzen, 1966; Schupp, 1986; Madden & Young, 1992; Rocha & Bergallo, 1992; Fiala et al., 1994; Chamberlain & Holland, 2009; Rosumek et al., 2009). Other plants produce unconcealed nectar, in floral or extrafloral nectaries, to attract defending wasps (Domínguez et al., 1989; Cuautle & Rico-Gray, 2003). Wasps visit the plant to feed on the nectar and, whilst doing so, remove any herbivores that they encounter (Domínguez et al., 1989).
In all known cases, plants that rely on these mutu-alistic relationships do not begin their lives with the necessary indirect defences to acquire or maintain them. Instead, they must do so during their ontogeny (Del Val & Dirzo, 2003). The absence of indirect defences during the early stages of plant growth has led to the suggestion that young (uncolonized) plants may be more reliant upon direct defences than older (colonized) ones (Collantes, Gianoli & Niemeyer, 1997; Heil & McKey, 2003).
Studies investigating the anti-herbivore defences of uncolonized myrmecophytes have varied widely in their support of the hypothesis that uncolo-nized young plants rely more upon direct defences than colonized ones. Nomura, Itioka & Murase (2001) showed that, in captivity, generalist herbivores ate
more and survived better on leaves clipped from plants after the initiation of symbiosis with ants than from plants before the initiation of this symbiosis in three Macaranga tree species. In contrast, Trager & Bruna (2006) showed that the specialist beetle Cop-tocycla leprosa fed equally on leaves from 1- and 5-year-old individuals of Cordia alliodora during laboratory trials (see also Del Val & Dirzo, 2003).
Furthermore, when comparing investment in direct defences between congeneric species with and without indirect defences, different studies have reached different conclusions. While some studies found sig-nificant differences in the amount of direct defences used by congeneric species with and without indirect defences (e.g. Rehr, Feeny & Janzen, 1973; Heil, Stae-helin & McKey, 2000; Eck et al., 2001), others have found no such differences (Steward & Keeler, 1988; Heil et al., 2002; Rudgers, Strauss & Wendel, 2004). Clearly, the lack of congruence among studies high-lights a need for studies that comprehensively control for several confounding factors.
Given the discordant results of past studies, we wanted to examine if two indicators of plant invest-ment in direct defences, herbivory and palatability (see Del Val & Dirzo, 2003), differ between plants with and without indirect defences. For this purpose, we compared the standing herbivory and palatability of two congeneric species with and without indirect defences at two ontogenetic stages: before and after the onset of indirect defences. We predicted that her-bivores would prefer to feed on species and stages with indirect or biotic defences.
MATERIAL AND METHODS STUDY SITE
We conducted this study at the Chamela Biological Station, within the Chamela–Cuixmala Biosphere Reserve (19°30′N, 105°03′W, Jalisco, Mexico), where the predominant vegetation is seasonally dry tropical forest. The area is characterized by its extreme seasonality. Average annual temperature is 25 °C and average annual precipitation is 750 mm, with a marked dry season from November to June (Bullock, 1986; Stoner, 2005).
The study was conducted during the rainy season of 2006 and field observations were carried out during the beginning of August, approximately 1 month after main leaf expansion occurred (Boege, 2004).
STUDY SPECIES
Cordia alliodora and Cordia elaeagnoides (Boragi-naceae) are widespread in the Neotropical Region. Cordia alliodora presents traits of a myrmecophyte and it is defended by ants (Tillberg, 2004; R. Dirzo,
pers. observ.). Cordia elaeagnoides does not have any apparent indirect defences. The domatia of C. alliodora are naturally hollow swellings produced at most branch nodes (Trager & Bruna, 2006). These nodes may be inhabited by different ant species in different parts of the plant’s range (Tillberg, 2004), but Azteca species are the most common ant partners of C. alliodora throughout the plant’s geographical range (Longino, 1996). In the case of C. alliodora at Chamela, the age at which the mutualistic interaction with Azteca ants begins differs among individuals, but ant colonization can be readily assessed by direct inspection. The approximate minimum size at which C. alliodora individuals recruit defenders is 65 cm overall height (R. Dirzo, pers. observ.).
Of the two selected Euphorbiaceae, Croton subero-sus has indirect defences whereas Croton pseud-oniveus does not (R. Dirzo, unpubl. data). Croton suberosus is distributed on the lowlands of Jalisco, Michoacán, Guerrero and Oaxaca, on the Pacific coast of Mexico. The floral nectar of C. suberosus attracts wasps which prey on herbivores (Domínguez et al., 1989). Although larvae of Hypercombe sp. (Lepi-doptera: Arctiidae) readily consume leaves of C. sub-erosus in laboratory test trials, larvae experimentally placed on leaves of C. suberosus plants in the field are readily killed and removed by wasps (Domínguez et al., 1989).
We sampled an area of forest approximately 1.5 km2. Sampled plants were interspersed with plants from other species. For the purpose of this experiment, we used the following operational defini-tions of young (uncolonized) and old (colonized) indi-viduals. In Croton suberosus (a shrub), we selected plants less than 30 cm tall for the young group and plants taller than 60 cm for the old group. Old Croton suberosus plants ranged between 0.6 and 3 m and they all had flowers. For the other three (tree) species, we selected plants less than 50 cm for the young group and plants taller than 1.5 m for the old group. Above this threshold, we estimated the herbivory of the smallest individuals we found. To avoid the risk that some very young plants of Cordia alliodora might have been colonized by ants, we verified that none of the young plants selected for this study had developed such an association. None of the young individuals of Croton suberosus had flowers at the time of sampling, whereas flowers were present on all old sampled individuals of this species.
PREFERENCE EXPERIMENT
We chose caterpillars of Hypercombe sp. for the experimental work because this species is a generalist feeder, which feeds regularly on all four plant study species (Domínguez et al., 1989). We collected
cater-pillars from the field each morning and used them for preference trials the same day. We kept caterpillars in plastic containers, covered with mosquito netting for ventilation, with wet cotton wool for moisture and a mixture of native plants, not including the study species, to feed on. Six hours before the onset of the experiment, we replaced the plant material in the containers with lettuce (Lactuca sativa L.) to mini-mize the effect of previous diet. After 2 h, we removed the lettuce and starved caterpillars for 4 h. This pro-cedure standardized caterpillar hunger (see Dirzo, 1980).
We determined each caterpillar’s preference of two plants, each plant from a different group. We used only mature leaves (i.e. in positions 3–4 of the phy-lotaxis) for the trials. We cut out two 4-cm2 pieces from the mature leaves of each plant and placed the four squares along the wall of a plastic container, equally spaced. We arranged the leaf pieces so that the two squares from the same plant were across from each other. We placed one caterpillar in the centre of the container and removed it after 5 h of feeding. At the end of each trial, we photographed the remaining leaf material beside a 5-cm scale bar and determined the area consumed by the caterpillar using IMAGE J (Rasband, 2003). This procedure was used for two experiments. In the first one, we tested whether caterpillars preferred to eat species with indirect defences to species without them; the second tested whether caterpillars preferred the defended (old) ontogenetic stage to the undefended (young) one.
To test whether caterpillars preferred to eat species with indirect defences to species without them, we presented caterpillars with leaf material from two individuals from the same genus (Cordia or Croton) and ontogenetic stage (young or old) but of different species (one with, the other without indirect defences). We calculated caterpillar preference by estimating the difference in leaf area consumed between species with and without indirect defences. We replicated the trials 15 times for each species pair to assess interspecific differences in direct defences. Preference for the species with indirect defences was analysed with a full factorial ANOVA with genus and ontogenetic stage as fixed factors. Because the variance of the residuals differed between groups, we rank transformed our dependent variable to fulfil the assumptions of the ANOVA model. This transformation is correct and powerful for 2¥ 2 factorial designs like the one used for this experiment (Thomson, 1991).
To test whether caterpillars preferred the defended (old) ontogenetic stage to the undefended (young) one, we presented caterpillars with leaf material from two individuals belonging to the same species, but in different ontogenetic stages (one young individual,
one old individual). We calculated the preference of caterpillars for old individuals as the area of leaf material consumed of the old individual minus the amount consumed of the young individual. We con-ducted 15 replicates for each choice test comparing young and old plants within each of the four species. Preference for old individuals was analysed with a full factorial ANOVA, with genus and presence of indirect defences as fixed factors.
FIELD OBSERVATIONS
To test if caterpillar preference explains herbivory in natural conditions, we conducted field observations of herbivory. We measured herbivory level and plant size (height, trunk diameter and number of leaves) of 15 individuals of each study species. In order to estimate the degree of herbivory in individual plants, we visually categorized damage using the method described in Dirzo & Domínguez (1995). We first determined a level of damage as a percentage for each leaf and then assigned leaves to different categories 0, 0–20, 20–40, 40–60, 60–80, 80–100 and 100. These seven categories were given a label of i= 0 to i = 6 and the expected percentage of leaf consumed (ci) for each
category was the midpoint of its range. We denoted the number of leaves in category i as ni, and the
number of leaves sampled as N. With these conven-tions, the percentage of leaf material consumed for an entire plant can be approximated by H= S[(ni ci)]/N.
We calculated H on the basis of all the leaves in a plant except for large trees, where we measured the herbivory of the leaves on three branches chosen at random.
We tested for effects of ontogenetic class (young or old), genus (Cordia or Croton) and indirect defences (presence or absence) and their interactions on our measure of herbivory with a full factorial ANOVA. Initially, we analysed the data including height, trunk diameter and number of leaves as covariates, but we removed them from the analysis because they did not have any significant effect on our measure of her-bivory. Because the residuals of the model were not normally distributed, we used Monte Carlo simula-tions to calculate empirical P-values for this analysis (Davison & Hinkley, 2006). We ran a total of 1999 Monte Carlo simulations using the statistical package R, version 2.8.1 (R Development Core Team, 2008).
RESULTS PREFERENCE EXPERIMENT
Caterpillars showed a clear preference for species with indirect defences. Regardless of the genus and ontogenetic stage, caterpillars consumed 4.98 ± 2.39 cm2(mean ± SE) more leaf material from
species with indirect, biotic defences than from species without them (F1,56= 267.07; P< 0.01; Fig. 1). There was a significant effect of the age class–genus interaction on the preference for species with indirect defences (F1,56= 7.17, P< 0.01): the preference for young Cordia alliodora over young C. elaegnoides was smaller than for the other groups (Fig. 1).
Overall, there was a marginally significant effect of within-species ontogenetic stage on the amount of leaf material the caterpillars consumed (F1,56= 4.28; P= 0.04), but there was a significant interaction between ontogenetic stage and plant defensive strat-egy (F1,56= 15.58, P< 0.01). Caterpillars preferred leaves from older individuals of plant species without them (Fig. 2). Neither genus (F1,56= 0.32, P = 0.57) nor its interaction with presence of indirect defences (F1,56= 0.69, P = 0.41) had an effect on the preference for old individuals.
FIELD OBSERVATIONS
Only the presence or absence of indirect defences had a statistically significant effect on natural levels of herbivory (Table 1): species with indirect defences experienced less herbivory in the field than species without indirect defences (Fig. 3, P= 0.03). None of the other factors, or their interactions, had significant effects, with the exception of the interaction between presence or absence of indirect defences and ontoge-netic stage that was also significant (P= 0.04). Although, overall, plants with indirect defences suf-fered less herbivory than plants without indirect defences, the difference was marked mainly in old plants and it was much less so in young plants (Fig. 3).
Figure 1. Preference (mean ± 95% CI) for species with
indirect defences in the comparison with vs. without indi-rect defences for the two genera (Croton and Cordia) and the two ontogenetic stages (adult and young) of this study.
DISCUSSION
Our results are consistent with the assumption that defensive traits are costly. Because plants have a finite amount of resources, they face a trade-off between investing in different defensive strategies (Janzen, 1966; Heil et al., 1999, 2000; Nomura, Itioka & Itino, 2000; Eck et al., 2001). This is not a new idea, but it remains controversial. Agrawal (2006) pointed out two major difficulties when comparing defensive strategies between species with and without indirect defences. First, some studies have failed to take into account phylogenetic relationships when comparing between species with and without indirect defences. Second, many studies have concentrated on only searching for univariate trade-offs. For instance, Heil et al.
(2000) compared chitinase activity between leaves of a myrmecophytic species and three other non-myrmecophytic species (see also Rehr et al., 1973; Eck et al., 2001). These studies implicitly assumed that the focal chemical (or the family of chemicals) constitutes a defence in itself. However, plants relying on direct defences to discourage herbivore attacks mount complex defensive systems combining different defen-sive traits. Even if there is a trade-off between invest-ing in indirect or direct defences, there is no reason to expect a negative correlation between investment in indirect defences and any one particular component of the direct defence strategy.
In one of the most comprehensive studies on trade-offs between investing in indirect and direct defences, Heil et al. (2002) found no clear evidence that plants with indirect defences invest less in chemical defences. However, their results were inconclusive
Figure 2. Preference (mean ± 95% CI) for the adult stage
in the comparison adult vs. young stage for species with and without indirect defences and for the two genera (Croton and Cordia). The dashed line indicates indiffer-ence between the two ontogenetic stages. Positive values indicate a preference for old individuals and negative values a preference for young individuals.
Table 1. Results of ANOVA testing the effect of
ontoge-netic stage (colonized vs. uncolonized), genus (Cordia vs. Croton) and indirect defences (present vs. absent) on the herbivory index
Source of variation d.f. F P
Genus 1 0.003 0.956
Indirect defences 1 4.840 0.026 Stage class 1 0.044 0.830 Genus¥ indirect defences 1 0.532 0.478 Genus¥ stage class 1 0.887 0.363 Indirect defences¥ stage class 1 3.851 0.038 Genus¥ indirect defences ¥ stage class 1 1.398 0.243
Error 112
Figure 3. Mean herbivory indices of species with and without indirect defences divided by stage classes within the genera (A) Cordia and (B) Croton. Bars denote standard errors.
because of the use of an artificial diet, which lacked physical defences, was rich in proteins and had con-centrations of chemical defences much lower than those found in the actual leaves. This is critical because the effect of toxin concentration in the diet on consumption and growth is non-linear (Yu, 1987; Wadleigh & Yu, 1988; Johnson, 1999) and depends on the protein-to-carbohydrate ratio of the food (Simpson & Raubenheimer, 2001).
The results of our within-species comparisons (Fig. 2) are consistent with the requirement for plants to trade-off between investing in indirect and direct defences. Plants lacking indirect defences did not show a significant difference in their palatability as they matured, but for species with indirect defences caterpillars preferred plants that had already acquired the indirect defences. That individuals lose or reduce their direct defences during the adult stages is an indication that these defences are costly. This pattern fits the prediction of Boege & Marquis (2005), which states that there should be an increase in defence from the juvenile to the mature stage, and it highlights the importance of mechanistic constraints influencing ontogenetic variation in plant defences (Barton & Koricheva, 2010).
As in the present study, some previous studies have found that, in plant species with indirect defences as adults, investment in direct defences decreases when indirect defences become available (Nomura et al., 2001). This is not, however, a universal result. In some cases, the inconsistency in results can be explained on the basis of methodological differences, such as differences in the criteria defining the onto-genetic stages or use of specialist herbivores (Trager & Bruna, 2006), but inconsistencies between experi-mental results cannot always be explained purely on methodological grounds (e.g. Del Val & Dirzo, 2003). It is therefore important to note that there is no single general pattern in the ontogeny of plant defence (Barton & Koricheva, 2010) and that there is as yet no theory that adequately explains the variety of changes in plant defensive strategy deployed through ontogeny (Boege & Marquis, 2005; Barton & Koricheva, 2010).
Despite the caterpillar’s preference for species with indirect defences, standing herbivory in the field was higher in plants without indirect defences than with indirect defences. This pattern was particularly clear among old plants (Fig. 3). There are two likely expla-nations for this pattern. Firstly, the low herbivory level of potentially preferred plants (those with indi-rect defences) could be a result of generalist herbi-vores being forced to feed on their unpreferred food sources because of the high predation risk (in Croton suberosus) or caterpillar attack and subsequent falling to the ground (in Cordia alliodora) in their
preferred food sources (Lima & Dill, 1990). Secondly, the direct defences of plant species without indirect defences may repel generalist herbivores in the field, and the high herbivory we observed could attest to the presence of specialised herbivores well equipped to deal with direct plant defences. Either way, plants with indirect defences proved to be better protected against herbivores than plants without them (see also Dyer et al., 2004).
Finally, we suggest that trade-offs between alterna-tive defensive traits must be one of the basic assump-tions in any attempt to explain the evolution of defence strategies. That these trade-offs exist is tes-tament to the fact that not all plant species rely on indirect defences to remove herbivores. If, in accor-dance with the results of our field observations, plants with indirect defences suffer less herbivory than plants without them, why, then, do many plant species not use indirect defences? We posit two pos-sible explanations. First, that phylogenetic inertia prevents some plant species from evolving indirect defences or, second, that indirect defences are costly, so that the cost of recruiting herbivore predators balances the benefit obtained from them. Given the broad taxonomic distribution of indirect defences, it seems unlikely that phylogenetic constraints prevent other plants from adopting this defence strategy. It seems reasonable to suggest that the presence or absence of indirect defences can be explained with a cost-benefit framework. Clearly, this is an aspect that warrants further research. We conclude that indirect defences are an effective defence against herbivores and that their recruitment allows plants to reduce investment in other defence mechanisms, supporting the notion of plants trading-off between defensive strategies.
ACKNOWLEDGEMENTS
Thanks to Irene López for help with the fieldwork and to Oriol Verdeny for statistical help. Financial support to A.L.L. came from a research grant of the British Ecological Society (SEPG, reference no. 755– 877). Elizabeth Pringle and Kate Umbers read a previous draft and made valuable comments. The Chamela Tropical Research Station provided logisti-cal support to undertake field and laboratory work. Thanks to two anonymous reviewers for their helpful comments on an earlier version of this paper.
REFERENCES
Agrawal AA. 2006. Macroevolution of plant defense strategies. Trends in Ecology and Evolution 22: 103–109.
Barton KE, Koricheva J. 2010. The ontogeny of plant defense and herbivory: characterizing general patterns using meta-analysis. American Naturalist 175: 481–493. Boege K. 2004. Induced responses in three tropical dry forest
plants species – direct and indirect effects on herbivory.
Oikos 107: 541–548.
Boege K, Marquis RJ. 2005. Facing herbivory as you grow up: the ontogeny of resistance in plants. Trends in Ecology
and Evolution 20: 441–448.
Bryant JP, Reichardt PB, Clausen TP, Provenza FD, Kuropat PJ. 1992. Woody plant–mammal interactions. In: Rosenthal GA, Berenbaum MR, eds. Herbivores: their
inter-actions with plant secondary metabolites. San Diego, CA:
Academic Press, 344–371.
Bullock SH. 1986. Climate of Chamela, Jalisco, and trends in the south coastal region of Mexico. Archives for Meteorology,
Geophysics and Bioclimatology Series B 36: 297–316.
Chamberlain SA, Holland JN. 2009. Quantitative synthesis of context dependency in ant–plant protection mutualisms.
Ecology 90: 2384–2392.
Collantes HG, Gianoli E, Niemeyer HM. 1997. Effect of defoliation on the patterns of allocation of a hydroxamic acid in rye (Secale cereale). Environmental and Experimental
Botany 38: 231–235.
Cuautle M, Rico-Gray V. 2003. The effect of wasps and ants on the reproductive success of extrafloral nectaried plant
Turnera ulmifolia (Turneraceae). Functional Ecology 17:
417–423.
Davison AC, Hinkley D. 2006. Bootstrap methods and
their applications, 8th edn. Cambridge: Cambridge Series
in Statistical and Probabilistic Mathematics.
Del Val E, Dirzo R. 2003. Does ontogeny cause changes in the defensive strategies of the myrmecophyte Cecropia
peltata? Plant Ecology 169: 35–41.
Del Val E, Dirzo R. 2004. Myrmecophily: plants with their own army. Interciencia 29: 673–677.
Di Giusto B, Anstett MC, Dounias E, McKey DB. 2001. Variation in the effectiveness of biotic defence: the case of an opportunistic ant–plant protection mutualism. Oecologia 129: 367–375.
Dirzo R. 1980. Experimental studies on slug–plant interac-tions. I. The acceptability of thirty plant species to the slug
Agriolimax caruanae. Journal of Ecology 68: 981–998.
Dirzo R, Domínguez CA. 1995. Plant–herbivore interactions in Mesoamerican tropical dry forests. In: Bullock SH, Medina E, Mooney HA, eds. Seasonally dry tropical forests. Cambridge: Cambridge University Press, 304–325. Domínguez CA, Dirzo R, Bullock SH. 1989. On the
func-tion of floral nectar in Croton suberosus (Euphorbiaceae).
Oikos 56: 109–114.
Dyer LA, Letourneau DK, Dodson CD, Tobler MA, Stireman JO, Hsu A. 2004. Ecological causes and conse-quences of variation in defensive chemistry of a Neotropical shrub. Ecology 85: 2795–2803.
Eck G, Fiala B, Linsenmair KE, Hashim RB, Proksch P. 2001. Trade off between chemical and biotic anti-herbivore defense in the Southeast Asian plant genus Macaranga.
Journal of Chemical Ecology 27: 1979–1996.
Fiala B, Grunsky H, Maschwitz U, Linsenmair KE. 1994. Diversity of ant–plant interactions – protective efficacy in
Macaranga species with different degrees of ant association. Oecologia 97: 186–192.
Fiala B, Maschwitz U. 1992. Food bodies and their signifi-cance for obligate ant-association in the tree genus
Macar-anga (Euphorbiaceae). Botanical Journal of the Linnean Society 110: 61–75.
Gaume L, McKey D, Terrin S. 1998. Ant–plant–homopteran mutualism: how the third partner affects the interaction between a plant-specialist ant and its myrmecophyte host.
Proceedings of the Royal Entomological Society of London. Series B 265: 569–575.
Goralka RJL, Langenheim JH. 1996. Implications of foliar monoterpenoid variation among ontogenetic stages of California bay tree (Umbellularia californica) for deer herbivory. Biochemical Systematics and Ecology 24: 13–23. Goralka RJL, Schumaker MA, Langenheim JH. 1996. Variation in chemical and physical properties during leaf development in California bay tree (Umbellularia
califor-nica): predictors regarding palatability for deer. Biochemi-cal Systematics and Ecology 24: 93–103.
Gowda JH, Albrectsen BR, Ball JP, Sjöberg M, Palo RT. 2003. Spines as a mechanical defence: the effects of fertiliser treatment on juvenile Acacia tortilis plants.
International Journal of Ecology 24: 1–4.
Heil M, Delsinne T, Hilpert A, Schürkens S, Andary C, Linsenmair KE, Sousa SM, McKey D. 2002. Reduced chemical defence in ant-plants? A critical re-evaluation of a widely accepted hypothesis. Oikos 99: 457–468.
Heil M, Fiala B, Boller T, Linsenmair KE. 1999. Reduced chitinase activities in ant plants of the genus Macaranga.
Naturwissenschaften 86: 146–149.
Heil M, McKey D. 2003. Protective ant–plant interactions as model systems in ecological and evolutionary research.
Annual Review of Ecology, Evolution, and Systematics 34:
425–453.
Heil M, Staehelin C, McKey D. 2000. Low chitinase activity in Acacia myrmecophytes: a potential trade-off between biotic and chemical defences? Naturwissenschaften 87: 555– 558.
Herms DA, Mattson WJ. 1992. The dilemma of plants: to grow or defend. Quarterly Review of Biology 67: 283–335. Janzen DH. 1966. Coevolution of mutualism between ants
and Acacias in Central America. Evolution 20: 249–275. Janzen DH. 1969. Allelopathy by mirmecophytes: the ant
Azteca as an allelopathic agent of Cecropia. Ecology 50:
147–153.
Johnson KS. 1999. Comparative detoxification of plant (Magnolia virginiana) allelochemicals by generalist and specialist Saturniid silkmoths. Journal of Chemical Ecology 25: 253–269.
Kost C, Heil M. 2005. Increased availability of extrafloral nectar reduces herbivory in Lima bean plants (Phaseolus
lunatus, Fabaceae). Basic and Applied Ecology 6: 237–248.
Lima SL, Dill LM. 1990. Behavioral decisions made under the risk of predation: a review and prospectus. Canadian
Longino JT. 1996. Taxonomic characterization of some live-stem inhabiting Azteca (Hymenoptera: Formicidae) in Costa Rica, with special reference to the ants of Cordia (Boraginaceae) and Triplaris (Polygonaceae). Journal of
Hymenoptera Research 5: 131–156.
Madden D, Young TP. 1992. Symbiotic ants as an alterna-tive defense against giraffe herbivory in spinescent Acacia
drepanolobium. Oecologia 91: 235–238.
Nomura M, Itioka T, Itino T. 2000. Variations in abiotic defense within myrmecophytic and non-myrmecophytic species of Macaranga in a Bornean dipterocarp forest.
Ecological Research 15: 1–11.
Nomura M, Itioka T, Murase K. 2001. Non-ant antiherbi-vore defenses before plant–ant colonization in Macaranga
myrmecophytes. Population Ecology 43: 207–212.
R Development Core Team. 2008. R: a language and
envi-ronment for statistical computing. Vienna, Austria: R
Foun-dation for Statistical Computing.
Rasband W. 2003. Image J. 1.30. National Institutes of Health. Available at http://rsb.info.nih.gov/ij/
Rehr SS, Feeny PP, Janzen DH. 1973. Chemical defence in Central American non-ant Acacias. Journal of Animal
Ecology 42: 405–416.
Rocha CFD, Bergallo HG. 1992. Bigger ant colonies reduce herbivory and herbivore residence time on leaves of an ant-plant Azteca muelleri vs Coelomera ruficornis on
Cecropia pachystachya. Oecologia 91: 249–252.
Rohde S, Molis M, Wahl M. 2004. Regulation of anti-herbivore defence by Fucus vesiculosus in response to various cues. Journal of Ecology 92: 1011–1018.
Rosumek FB, Silveira FAO, Neves FS, Barbosa NPU, Diniz L, Oki Y, Pezzini F, Fernandes GW, Cornelissen T. 2009. Ants on plants: a meta-analysis of the role of ants as plant biotic defenses. Oecologia 160: 537–549.
Rudgers JA, Strauss SY, Wendel JF. 2004. Trade-offs among anti-herbivore resistance traits: insights from Gos-sypieae (Malvaceae). American Journal of Botany 91: 871– 880.
Schemske DW. 1980. The evolutionary significance of extra-floral nectar production by Costus woodsonii (Zingiberaceae) – an experimental analysis of ant protection. Journal of
Ecology 68: 959–967.
Schupp EW. 1986. Azteca protection of Cecropia:
ant occupation benefits juvenile trees. Oecologia 70: 379– 385.
Simpson SJ, Raubenheimer D. 2001. The geometric analy-sis of nutrient–allelochemical interactions: a case study using locusts. Ecology 82: 422–439.
Steward JL, Keeler KH. 1988. Are there trade-offs among antiherbivore defenses in Ipomoea (Convolvulaceae)? Oikos 53: 79–68.
Stoner KE. 2005. Phyllostomid bat community structure and abundance in two contrasting tropical dry forests.
Biotropica 37: 591–599.
Thomson GL. 1991. A note on the rank transform for interactions. Biometrika 78: 697–701.
Tillberg CV. 2004. Friend or foe? A behavioral and stable isotopic investigation of an ant–plant symbiosis. Oecologia 140: 506–515.
Trager MD, Bruna EM. 2006. Effects of plant age, experi-mental nutrient addition and ant occupancy on herbivory in a neotropical myrmecophyte. Journal of Ecology 94: 1156– 1163.
Wadleigh RW, Yu SJ. 1988. Detoxification of isothiocyanate allelochemicals by glutathione transferase in three lepi-dopterous species. Journal of Chemical Ecology 14: 1279– 1284.
Webber BL, Moog J, Curtis ASO, Woodrow IE. 2007. The diversity of ant–plant interactions in the rainforest under-storey tree, Ryparosa (Achariaceae): food bodies, domatia, prostomata and hemipteran trophobionts. Botanical Journal of the Linnean Society 254: 353–371.
Yu SJ. 1987. Microsomal oxidation of allelochemicals in generalist (Spodoptera frugiperda) and semispecialist (Anticarsia gemmatalis). Journal of Chemical Ecology 13: 423–436.
